Artificial neural networks as a tool for mineral potential mapping with GIS

Abstract

Abstract. A back-propagation artificial neural network (ANN) model is proposed to discriminate zones of high mineral potential in the Rodalquilar gold field, south-east Spain, using remote sensing and mineral exploration data stored in a GIS database. A neural network model with three hidden units was selected by means of the k-fold cross-validation method. The trained network estimated a gold potential map efficiently, indicating that both previously known and unknown potentially mineralized areas can be detected. These initial results suggest that ANN can be an effective tool for mineral exploration spatial data modelling.

Análisis e integración de datos espaciales en investigación de recursos geológicos mediante Sistemas de Información Geográfica

“…En el primer caso, los parámetros son estimados sobre la base de la opinión de un experto en el tema, y en el segundo son obtenidos del análisis de las relaciones espaciales entre las capas independientes y la capa dependiente. Los modelos basados en el conocimiento hacen uso de funciones de integración tales como la lógica booleana, la suma ponderada o la lógica difusa, mientras que los modelos basados en los datos utilizan típicamente funciones como la regresión múltiple, el análisis discriminante, métodos probabilísticos bayesianos o incluso redes neuronales (Bonham-Carter et al, 1989;Agterberg et al, 1993;Bonham-Carter, 1994;Rigol-Sanchez et al, 2003). Los modelos basados en la suma ponderada y en la regresión múltiple son, debido a sus características, dos de los métodos más utilizados.…”

Gold resources potential assessment in eastern Kunlun Mountains of China combining weights-of-evidence model with GIS spatial analysis technique

“…In recent years, there have been growing literatures on the application of weights-of-evidence model for mineral exploration (Carranza, 2004;Daneshfar et al, 2006), and the model has also been applied in other fields such as animal habitats (Romero and Luque, 2006), geologic hazards (Zahiri et al, 2006;Neuhäuser and Terhorst, 2007;Song et al, 2008), groundwater resources (Cheng, 2004;Corsini et al, 2009) and hydrology pollution (Masetti et al, 2007). In the last few years, some other spatial statistical models also have been used in mineral resources assessment, such as logistic regression model (Agterberg et al, 1993;Sahoo and Pandala, 1999;Carranza and Hale, 2001), fuzzy logic model (Luo and Dimitrakopoulos, 2003), artificial neural networks model (Koike et al, 2002;Rigol-Sanchez et al, 2003;Nykänen, 2008). Compared with other models, the weights-of-evidence model has some advantages in assessing mineral resources: 1) the weights are relatively easy to be interpreted, they can be confirmed independently, and the favorable targets can be identified easily from the posterior probability or the sum of weights for visual analysis; 2) using proximity analysis to obtain optimal cut-offs, the method provides better estimates for contrast, studentised contrast and buffer size; and 3) it also can be used to capture suitable fuzzy membership (Cheng and Agterberg, 1998;Porwal et al, 2003).…”

A predictive GIS model for mapping potential manganese mineralization in western Guangxi and southeastern Yunnan area, China

“…The weights-of-evidence methodology was implemented in a GIS framework to quantify the spatial correlations between the deposit recognition criteria and the known mineral occurrences and predict prospective areas [9]. A back-propagation artificial neural network (ANN) model was proposed to discriminate zones of high mineral potential in the Rodalquilar gold field, south-east Spain [10]. The Chinese scholars also developed a series of software about mineral resources prediction and evaluation based on GIS platform, such as mineral resources assessment system (MRAS) [II,12], mineral ore resources perspective and assessment system (MORPAS) [13,14], GeoDAS [15,16], KCYC [17].…”